Tree Species Composition and Structureofa Coastal Hill Forest in Pulau Pangkor , Malaysia

Tree species composition and structure of a coastal hill forest in Sungai Pinang Permanent Forest Reserve in Pulau Pangkor at Perak was studied based on the data from five 1-hectare plots. All stems with a diameter at breast height (dbh) ≥ 5cm enumerated, identified and their height weremeasured. The mean stem density was 659 (stems ha), and a total of 45 families, 102 genera, 181 species, 3319 individuals were presented in the five 1-ha plot; this comprises 6.3% species, 19.2% genera and 42% families of the total tree taxa found in Peninsular Malaysia. Based on IVI, Vatica maingayi,Xanthophyllum affine,Vatica pauciflorawere the most dominant tree species. The families of Dipterocarpaceae, Polygalaceae,Guttiferae and Myrtaceae were among the most important families with high FIV. Based on dbh class, nearly 50% of the trees were distributed in the lower diameter class(5-15cm). According to heightlife-form categories, 53% of stems were belonged to small trees (7-15m). Species accumulation curve showedan asymptote at the third hectare.


Introduction
Tropical rainforest in five major regions (Corlett and Primack 2006), have the most diverse terrestrial ecosystems (Turner 2001) and extremely important for conserving biodiversity for the local and regional development and also for sustainable use of forests (Neshim et al.2009).
Malaysian forests harbour a very large portion of diversity and forests have an important role in the country socio-economic development and environmental conservation (Nizam Khan and Zakaria 2005).This diversity is under great threat with total forested area in Peninsular Malaysia with 44.7% of its land area (Abdul Rashid 2005).There are 16 forest types based on altitude and soil types and distributed from the coastal lines to the mountain (Whitmore and Sayar 1992).
The flora of Peninsular Malaysia contains the floristic elements of the Sunda Shelf and also of the mainland Asiatic species from seasonal climates (Saw and Chung 2004) and has been studied by several botanists.The first effort was done by Ridley (1922Ridley ( -1925) ) by Flora of the Malay Peninsula, while Whitmore andNg (1972-1989) prepared the forest trees of Peninsular Malaysia in the Tree Flora of Malaya (Vol. 1 to IV).The current checklist catalogues reveals, Peninsular Malaysia has about 8893 taxa (species, subspecies and varieties) covering around 8200 native and 690 naturalised species (Sawet al. 2009) but the flora is not completely known and continuous botanical inventories is an on-going process.
Tree species constitute a large proportion of the flora of the Malay Peninsula.From 8096 species that comprise the native flora of the Malay Peninsula (Turner, 1997), 2830 species are trees (35%) in which are distributed in 106 families and 532 genera (Ng 1991).
Coastal hill forests are the climax vegetation.The main difference between the hill dipterocarp forests and lowland forests is the specific composition of the dominant of the upper strata of the vegetation.Relatively few studies have considered the plant diversity in a coastal hill forest.Symington (2004) recognized a floristic similarity between the costal hill and the hill dipterocarp forests.Abd.Rashid (2005) and Azhar Mudaet al. (2006) provided a list of plants found in coastal hill forest in Terengganu.But there is still insufficient information for Forestry Department of Malaysia to act upon especially for the consolidation of plant resources in coastal hill forest.This study focused on tree species richnessand evaluatesfloristic composition in a coastal hill forest.

Study area
This study was performed in Sungai Pinang Permanent Forest Reserve, Pulau Pangkor an island in Perak, north-west Peninsular Malaysia between 04°13.0'Nlatitude and 100° 33.0´E longitude (Figure 1).It is one of the famous and well known islands in Malaysia with area of eight square kilometres and classified as coastal hill forest with a high conservation.The climate is typically humid tropical and seasonal heavy rain (MOSTE 1997),February and March are the driest months, mean annual rainfall is 1820.23 mm (2000-2010).The highest mean temperature is in February to May (average 27.65 °C) and the minimum occurs during September to December (average 26.8 °C) (Meteorology Office, Sitiwan Station, Perak).The elevation of the study area ranges from 45to 350 m above the sea level.The texture of topsoil is mainly sandy loam.

Methods and data analysis
As 1 ha plots widely used and considered of appropriate size for studies in tropical forest, five plots of 1-ha (each 100 ×100m) were established and subdivided into 100 subplots of 10×10 m between 45 and 152 m above sea level.Each plot was systematically surveyed by enumerating, identifying and measuring all trees with height and diameter at breast height (DBH) ≥ 5cm.DBH was measured at 1.3 m above the ground unless the stem was swollen; measurements were taken at the nearest lower point where the stem was cylindrical.Based on Kitamura et al. (2005) four life-form categories including: arboreal shrub (SH<7), small tree (ST 7-15m), mid-sized tree (MT 15-30m), and tall tree (TT>30m) were determined.The inventories were made in December 2008 till October 2009.Five soil samples from each plot were collected to determine the soil texture.The specimens were fixed and kept in herbarium of Forestry Faculty, in Universiti Putra Malaysia and identified on the basis of regional flora.In addition, local names were recorded and the final verification was done at the Forestry Research Institute Malaysia (FRIM).
The importance value index (IVI) was assigned to describe the species composition of the plots and was calculated as the sum of the following three variables (Curtis and McIntosh, 1951): Relative dominance = (total basal area for a species/total basal area for all species)×100 Relative density = (number of individuals of a species/total number of individuals)×100 Relative frequency = (frequency of a species/sum frequencies of all species)×100 The family importance values (FIV) (Mori et al. 1983) were used to compare the relative contribution of each taxonomic family to forest species composition and was calculated as the sum of the following three variables: Relative diversity = (number of species in family/total number of species) × 100 Relative density = (number of individuals in family/total number of trees) × 100 Relative dominance = (basal area of family/total basal area) × 100 Also tree density for each plot (Tree density = Number of trees/area) calculated (Narayanan and Swarupanandan 1996).

Floristic composition
A total of 3319 individualstrees with dbh≥ 5 representing 181 species belong to 102 genera in 45 families were identified within five 1-ha plots of a coastal hill forest that is comprises 6.4% species, 19% genera and 42% families of the total tree taxa found in peninsular Malaysia.Vatica maingayi (IVI 300.46),Xanthophyllum affine (IVI 177.33),followed by Vatica pauciflora (IVI165.73)were dominant tree species (Table 1,2).The most speciose were Dipterocarpaceae having 13 species or 7.18% of total number of species,followed by Guttiferae (8 species or 4.42%), Euphorbiaceae (7 species or 3.87%) andPolygalaceae (5 species or 2.76%), while more than 23 families were singletons (represented by only one species).Table 3 contrasts the FIV for all families in the study area in which it was characterised by a rich value of 506.4 for Dipterocarpaceae.
The number of families, species numbers and individual numbers within five plots were significantly difference (P<0.001,P<0.0001, P<0.001, respectively).

DBH and Basal area
Species richness in terms of the number of species decreased with increased diameter classes.Trees within class 5-14.9 cm dbh constitute an average of about 54% of individuals and the distribution for the others declined dramatically in which 21% of individuals belonged to the next higher range (15-24.9cm) (Figure2).However several species (e.g.Shorea lumutensis, Xanthophyllum affine and Syzygium filiforme) attained dbh > 65 cm representing 2.5% of the total individuals.Figure 3 shows the distribution of basal area by dbh classes in five 1-ha plots.The total basal area for 3319 stemswas 28.17 m 2 ha -1 and Dipterocarpaceae with 744 individuals had the largest basal area.Table 4 shows the 20 largest basal areas by species in this study area.Mean DBH between five plots were significantly difference (P<0.0001).

Species accumulativecurve
Species accumulative curve reached an asymptote and flatten out, since the curve began to level off at third plot, it suggests our sampling area captured maximum proportion of the species richness (Figure 4).As tropical ecologists believe tree species richness reaches an asymptote at 1-3 ha (Tuomisto et al. 1995, Condit et al. 1996).

Discussion
Species richness is the most important characteristics of tropical rainforest ecosystem.Peninsular Malaysia has high diversity because it has been geologically stable from early Tertiary (Keng 1978) and the stable environment makes it easier for plants to survive (Ganesen 2008).The distribution of tree species in coastal hill has narrow ecological range.Many of lowland dipterocarp forest genera are represented but with different species and composition.Shorea curtisii for example is mainly limits to the ridges and upper slope areas (Ashton 1964, Saiful et al. 2008).In this area the ridges were often dominated by other non-dipterocarps such as Swintonia sp., Xanthophyllum sp. and Sarcotheca sp.It is notable that climax or late-successional species like Sarcotheca griffithii was the only non-dipterocarp species with high abundance in the last plot and Xanthphyllum affine was abundant and widespread across the gradient.
Microclimate is important because of its effects on species richness and ecosystem functions.In this area total vegetation cover was related on moisture and decreased with distance from the stream, which is likely to be caused by lower levels of soil moisture.Many of species are dependent on the higher amount of moisture, thesespecies are moisture obligates and mainly appeared in the first plot(Table 5).
Polygalaceae in the first plot was the dominant family with FIV 32.83 but Dipterocarpaceae was the dominant family with FIV 44, 77, 256 and 111 in plots 2 to 5, respectively.The importance of Dipterocarpaceae with 156 species (Turner 1997) and fourteenth most speciousin the flora as an emergent canopy tree in this area not only in terms of species richness but also in the proportional abundance of individuals is widely appreciated and remarkable.The extreme variability in stocking of economic species (large trees) was found on ridge-tops and slopes where the highest concentration of commercial timber tends to occur.According to Newbery et al. (1992), the regeneration of dipterocarps is unlikely to occur on slopes of 45 deg.or more.This observation is of immense importance to the Peninsular Malaysian hill forest especially by virtue of the fact that such terrain contributes 15-20% of the total land area in the hills whereas about 40% of the total land area of the Peninsular rises above 170 m and about 23%that of 300 m.Seedling recruitment is also poorer in the upper hill dipterocarp forests than in the steep lowlands or hill-ridge forests.Table 6 shows the comparison of coastal hill forest with lowland forest in other sites of Malaysia.In general, plant richness compared with lowland forest was low (Table 8 shows the distribution of some species along the hill).
In the third plot there was a sharp decline on the number of species and individuals.The main reasons were the abundance of stemless palm Eugeissona tristis(Bertam) on the ridges in which it made barrier for the successful establishment of natural regeneration, heavy felling of trees due to more exposed nature of the ridges resulting in gulley erosion and mass-flow of soil and third reason was high elevation (350 m a.s.l).The negative influence of elevation on diversity has been widely reported in studies (Gazol and Ibanez 2009).The ground flora of plots (except first plot) is extremely poor, due partly to the common occurrence of the palm Eugeissona triste.
The higher values of diversity found on slopes could be attributed to the bare soil surfaces created by leaf litter movement downward (Facelli and Pickett 1991a) and by a less crowded overstorey that enabled higher radiation on the soil.In the valley bottoms regeneration is unevenly distributed and most of families with one species appeared there, large woody species were also often comparatively poor and the forest was characterised by the richness in the ground flora and shrub layer (The details of all vascular plants publishedinSeptember 2011 in a different paper).
There are three types of plant distribution, random, clump, and uniform (Ludwig and Reynolds 1988).Majority of species were found to be distributed randomly in the plots.Some species like Bouea oppositifolia, Agrostistachys longifolia, Mallotus penangensis and Xerospermum noronhianumappeared clump distribution.Some tree species are very local in their distribution.For example, Pangkor Island is the only place where Shorea lumutensis occurs naturally in Peninsular Malaysia.
The frequency distribution in different size classes showed strong reverse-J, revealed a negative exponential relation in population size between the size classes (Figure2) (Seng et al. 2004).Pandy and Shukla (2003) stated reverse J distribution, indicative of uneven-aged stands among several dbh classes that correspond to the condition of natural forest and regeneration of trees and Rozas (2006) revealed this is in accordance with the idealized well-balanced population structure of the forests.
The maximum basal area in the first 1-ha plot belonged to Myrtaceae and in the other plots Dipterocapaceae was the dominant family.Dipterocarpaceae accounted for 45.6% of the total basal area that shows this family represented by large trees.There was a large gap between Dipterocarpaceae (BA 10.6 m 2 ) and the next most abundant family, Polygalaceae, (BA 2 m 2 ) (Table 6).According to Ramos and Amo (1992), a mature stand of primary dipterocarp forest can have an estimated basal area within the range of about 36-46 m 2 /ha being common of tropical rainforest areas.An unmanaged primary lowland forest in Peninsular Malaysia usually has a basal area about 32 m 2 /ha.Comparing these figures with those obtained from this study, it is clear that the basal area of the stand under study was lower than reported by others(Table 7).Differences in basal area may be attributed to altitudinal variations, species composition, age structure and successional stage of the forest (Swamy et al. 2000).
The investigation of endemism and the conservation status of the species are getting increasingly important with issues of plant conservation to be dealt under current forest management practices.The number of endemic trees totals 746 species which represents 26.4% of the total number of tree species (Ng et al. 1990).The endemic species accounted 24 (3.2%) (Appendix 1).According to IUCN red list of threatened species (2010) and (Ng et al.(2008), there were 13 critically endangered species, 18 lower risk, 6 endangered and 7 vulnerable species in this island.

Conclusion
Species richness is one measure of biodiversity and it is very important for ecosystem functioning, stability and integrity.Apart from the ecological reasons, maintaining high biodiversity in forest ecosystems has economic, spiritual, ethical, scientific and educational importance.This study helps to complete the description of biodiversity and forest structure and may be sued to interpret aspects of forests and population dynamics community.
The consensus has been that much more research is needed to provide directly relevant and applicable evidence for a better understanding of the hill coastal ecosystem, including the growth behaviour of the trees and the factors affecting it.
These kinds of studies are important, for conservation biodiversity, for local and regional development and also sustainable use of forest, e.g. by combining ecotourism and harvesting non-timber.Table 8.The distribution of some species on different parts of the hill in study area

Figure 1 .Figure 2 .
Figure 1.Map of Study area in Peninsular Malaysia

Table 1 .
The 20 most abundant species in Sungai Pinang Permanent Forest Reserve in Pulau Pangkor according to decreasing order of IVI RDE: Relative density, RDI: Relative diversity, RF: Relative frequency

Table 2 .
The 20 least abundant species in Sungai Pinang Permanent Forest Reserve in Pulau Pangkor according to increasing order of IVI

Table 3 .
Important families according to decreasing order of FIV of five 1-ha plots in Sungai Pinang Permanent

Table 4 .
Tree basal area for 20 species according increasing order of Basal area in Sungai Pinang Permanent Forest Reserve in Pulau Pangkor

Table 6 .
Tree basal area for 20 families according to increasing order Sungai Pinang Permanent ForestReserve in

Table 7 .
Comparison of species richness and basal area in five 1-ha inventory plots, Sungai Pinang Permanent Forest Reserve in Pulau Pangkor plots in Sungai Pinang Permanent Forest Reserve in Pulau Pangkor Appendix 1.Total list of tree species ≥5 cm dbh in five 1-ha plots, Sungai Pinang Permanent Forest Reserve in